How would you measure and display temperature? Please help.

Please help as I have the situation of needing to measure the temperature of fluids in a pipe and display those temperatures in a way that is visible from 15 to 20 feet away.

I am just learning about arduino and its capabilities and want to try it out. I am building a display for a trade show and would like to show the temperature change across a plate fin heat exchanger. I would like to broadcast the temperatures to a large screen or segmented LED.

My initial instinct was to use a thermocouple. I am sure I can get the thermocouple into the flow while finding a way to get the wire out of the pipe without spilling the contents of the system. Being a rookie at this, I want to hear how those who are experienced with arduino would do it. I am also concerned about converting the information into a visual display.

Thanks in advance.

Get a couple of these: https://www.sparkfun.com/products/13266

And break up the project into several steps.

  1. Get a temperature printed to the serial monitor.
  2. Decide what display you want to use
  3. Get something to display.
  4. Display your temperature on the display device.

This is all well within the Arduino's capabilities.

I thought I posted some questions about this in another Thread - but now I can't find them.

I think they included

  • Are both sides of the heat exchanger "wet?
  • Post a diagram of the system you are working with.
  • What temperatures do you need to measure?

Provide as much information as possible - especially links to datasheets for products you are using.

...R

You might not need to get the sensor wet, the outside of the pipe may suffice. If the pipe is plastic, you might insert a copper section. The pic shows the return pipe under the dial thermo with two thermowells, a homebrew dry one with DS18B20 above a commercial wet one. The feed pipe simply has a DS18B20 for Arduino and another sensor for the digital thermo underneath a bit of insulation in the short horizontal section. It seems to be just as effective as the thermowells. Arduino at the top right matches the red heat meter. At the bottom of the feed and return pipes, you can see PT500 sensors, laboriously installed, never used.

I would suggest a DS18B20 if your fluid is in range -25>125C. It is digital, accurate, easy to use, cheap, and well-supported on this forum.

http://www.hacktronics.com/Tutorials/arduino-1-wire-tutorial.html

For a trade show, you are going to have a lot of people looking at it and asking questions. Hacking together an Arduino for the display is a good idea but the temperature sensors have to be the actual ones that your customers will use. You're trying to demonstrate the performance of the heat exchanger, aren't you? Then you need to do it in a way that your customers will understand.

If this was just a home project then I would say just tape or glue a 5-cent sensor to the outside of the pipe.

For industrial/automotive/aerospace use, you are going to have specialised sensors that everyone uses. They may even have worse performance than the homemade sensor, but they will have a known performance that your customers are used to. With a sensor that they are comfortable with, they will be comfortable with your performance claims.

will this help?

https://www.youtube.com/watch?v=U9EIsX1C0rQ

I've used little under $1 thermocouples in many places, using the SteinHardt-Hart equation (think that's the right spelling) to covert resistance to temperature. They're durable, reliable, and accurate.

First off, welcome.

second, assuming you know all you need about selection of sensors, you came here looking for electronics, specificaly help with selectin of both the input devies and the jumbo displays that can be seen from 20 feet away.

how fast are you looking for parts ?

max6675 is a K type, serial input and resolves to 0.25 deg C (has output in either C or F )
costs less than $10 on e-bay.

I see it on e-bay, but that can take weeks to get. most come with K-type, but I would assume that for a trade show, you want T-Type for the sensitivity.

it is probably the easiest way to get readings into the system. least work on your part.

========================

use any board, an UNO or NANO would work fine.
what you want is not really that hard, any board has the capabilities.

if you just want to buy your way out…
visit Cross Roads Electronics

I rarely recomend places, but I believe Crossroads offers pre-made boards with the drivers you need and can drive your jumbo numbers without you having to solder.

============================

DISPLAYS;

I have to leave this to others to offer JUMBO displays, not so large as a high school score board…

the other reason crossroads came to mind is that he is one of the guys how knows where to get displays.

A TPIC6C595 can be used to drive large common anode displays.

https://www.sparkfun.com/products/13279 https://www.sparkfun.com/products/8530

Leo..

jrdoner: I've used little under $1 thermocouples in many places, using the SteinHardt-Hart equation (think that's the right spelling) to covert resistance to temperature. They're durable, reliable, and accurate.

No you didn't. thermocouples output a voltage and have an almost zero resistance. Each type of thermocouple has a large polynomial you use to convert the voltage to temperature. You used a thermistor - which might be an option, but not what the OP asked for.

KeithRB: No you didn't. thermocouples output a voltage and have an almost zero resistance. Each type of thermocouple has a large polynomial you use to convert the voltage to temperature. You used a thermistor - which might be an option, but not what the OP asked for.

Right-on, Keith,

Having done this for a Rotax engine dual exhaust gas temperature, I can add a little. First, the thermocouples MUST have a 'real' or 'simulated' Cold Junction ... that is, a reference point that is 0C... as one may expect, this is usually done in hardware/software on a customized chip: called a thermocouple interface (sometimes known as a transconductance amplifier.) Inside this chip is a custom uC and a lookup table for the specific type of thermocouple utilized (the metal combinations.)

A common thermocouple amplifier is: https://www.adafruit.com/product/269

NIST has the official ITS-90 documents: http://srdata.nist.gov/its90/download/download.html

My experience with thermistors indicates around +/- 2C can be obtained rather reliably... with precision 1% components, +/- 1C can be managed. My most recent wireless thermometer: ESP8266 Wireless Temperature

When measuring the outside temperature of a pipe, you have a number of negative issues you need to address: - a rather large delta-t associated with fluid temperature in the pipe before a corresponding change in the outside temperature - pipe outside temperature may be several degrees (or more) lower than the internal fluid temperature

The pipe material affects the above concerns ... while metal pipe conducts temperatures better, they also dissipate temperatures quicker and act as a large, long heatsink. Plastic materials are inferior to metal for heat conductance. The best method of determine if the outside pipe monitoring is appropriate is Empirical Testing - that is, experiment.

Ray

mrburnette: A common thermocouple amplifier is: https://www.adafruit.com/product/269

My experience with thermistors indicates around +/- 2C can be obtained rather reliably... with precision 1% components, +/- 1C can be managed. My most recent wireless thermometer: ESP8266 Wireless Temperature

great, I was just about to order some MX6675 until your post....

The MAX6675 has been discontinued by Maxim

MAX31855K seems to be the chip of the day

as for the 'accuracy' of a thermocouple. getting a number to decimal places is resolution, having it return to the exact same output with the exact same input is repeatability, then there is hysteresys....

I accept that you can be 2 deg off, but have great sensitivity and repeatability. What did you find as regards these things ? if your sensitivity and repeatability are great, then you can just shift the reading in software.

Dave

dave-in-nj:
<…> MAX31855K seems to be the chip of the day

as for the ‘accuracy’ of a thermocouple. getting a number to decimal places is resolution, having it return to the exact same output with the exact same input is repeatability, then there is hysteresys…

I accept that you can be 2 deg off, but have great sensitivity and repeatability. What did you find as regards these things ?
if your sensitivity and repeatability are great, then you can just shift the reading in software.

Dave

Agreed with s/w fine-tuning. But, in all honesty, I did not have anything in my lab accurate enough to qualify the build… Slushly ice water showed 0C and boiling water was 99C but my GPS shows my elevation at 1045 feet. However, I used tap-water so there were impurities. Since this need is EGT, decimal accuracy was not required.

I observed no detectable hysteresis.
The stainless EGT sensor (eBay sourced) are rather heavy (mass) so it does take a couple of seconds to stablize.

Ray

Crazy talk: for anyone using the internal 1.1V AD reference, you can achieve aroundv5% to 6% temp reading from a “T” thermocouple without an amplifier! See here

mrburnette: Crazy talk: for anyone using the internal 1.1V AD reference, you can achieve aroundv5% to 6% temp reading from a "T" thermocouple without an amplifier! See here

I was involved in commercial HVAC and when you see the names Honeywell, Johnson Controls, Siemens, etc, you do not think that 'casual engineering' carries the day.

I was doing industrial accuracy, 1/10 of a percent, they were doing 5-10%..

my request was because I am trying to figure the most rugged, low cost, reliable, temperature sensors, and 2-3 degrees is fine. 'calibration' of often putting all the sensors in the same duct, running the fan and setting them so they all read the same at some common point, like room temperature.

I am looking at RTD's for their simplicity and ease of use. the MAX6675 caught my eye as it offered no need for fancy math, was just a library function and was serial input. so my 5 temp sensors was no big deal.

Add a current sensor and such and it is simple enough for a NANO to handle.

Hi, WOW 13 posts, and the OP hasn't returned in three days.

  • What is the fluid you wish to measure the temperature of?
  • Is the fluid under considerable pressure?
  • What is the pipe material?
  • Do you need the temperature sensor to be in the fluid or on the pipe wall?
  • How long would the leads of the sensors be, between the arduino and the sensors?
  • Why does the display have to be visible from 15 to 20feet.
  • If you could move the display closer, would a smaller display be better.

Now we wait...

Thanks.. Tom.. :)

dave-in-nj:
<…>
my request was because I am trying to figure the most rugged, low cost, reliable, temperature sensors, and 2-3 degrees is fine. ‘calibration’ of often putting all the sensors in the same duct, running the fan and setting them so they all read the same at some common point, like room temperature.

I am looking at RTD’s for their simplicity and ease of use. the MAX6675 caught my eye as it offered no need for fancy math, was just a library function and was serial input. so my 5 temp sensors was no big deal.

Add a current sensor and such and it is simple enough for a NANO to handle.

Take a look at this article if you like on using an ESP8266 and it’s analog input. I used 1% resistor for the voltage divider and a quality Vishay thermistor. I do now have in the lab an NIST “traceable” digital thermometer (+/- 1C) that I have used with the ESP8266 temperature probe and I’m showing a +/- 1C in the 50F to 90F range … which is all I wanted for the air handler in the HVAC I am monitoring.

Ray

TomGeorge: Hi, WOW 13 posts, and the OP hasn't returned in three days.

  • What is the fluid you wish to measure the temperature of?
  • Is the fluid under considerable pressure?
  • What is the pipe material?
  • Do you need the temperature sensor to be in the fluid or on the pipe wall?
  • How long would the leads of the sensors be, between the arduino and the sensors?
  • Why does the display have to be visible from 15 to 20feet.
  • If you could move the display closer, would a smaller display be better.

Now we wait...

Thanks.. Tom.. :)

well he said that he can put the sensor into the flow. so we do not need to concern ourselves about pipe materials, flow rates or even pressures. we have to assume that going to a trade show, the gent knows how to pipe up a thermowell.

ALL he is asking for really, is how to display so people walking by his both can see the values from about 20 feet back.

not sure why of the 13 posts, no one is trying to address his actual question.

dave-in-nj:
well he said that he and put the sensor into the flow. so we do not need to concern ourselves about pipe materials, flow rates or even pressures. we have to assume that going to a trade show, the gent knows how to pipe up a thermowell.

ALL he is asking for really, is how to display so people walking by his both can see the values from about 20 feet back.
not sure why of the 13 posts, no one is trying to address his actual question.

In my case, because I did not read the whole of the Op’s #1 post! Many of us have been down the temperature measurement path before and are somewhat (in my case) myopic and perhaps even a bit biased on approach. Were i to do a demo, I’d use a PC with an external monitor/projector and use the ESP8266 as a webserver much as I did in the project I wrote and referenced in the previous post… Javascript and websockets can be easily implemented on an older notebook with a modern browser… no need for a fancy new notebook. And the ESP8266 NodeMCU was $3.10 and the other stuff about $1… project cost under $5.

Ray

Hi, To display temperature at that distance, use a servo and make a huge analog meter, bit of a show attraction at a trade show.

Or make a bit Persistance of Vision display, either a rotating LED string, or if you have the time a laser and scanning mirror, project figures up on wall.

Tom.. :)

mrburnette:
Take a look at this article if you like on using an ESP8266 and it’s analog input.

thanks, GREAT link, but I feel like I am hi-jacking this thread.

I started a new thread and will only respond to the OP’s comments on here.

Dave